Long-term safety of human retinal progenitor cell transplantation in retinitis pigmentosa patients

Study design and ethics

The animal study was approved by the Laboratory Animal Welfare and Ethics Committee of the Third Military Medical University, and the clinical trial was approved by the Medical Ethics Committee of Southwest Hospital, the Third Military Medical University. We conducted a phase I clinical trial assessing the safety of RPC transplantation into RP patients over a 24-month follow-up period. This trial was conducted in the Southwest Hospital, China. Recruitment of patients started in May 2008 and the study was completed in September 2013. The research adhered to the principles of the Declaration of Helsinki, and written informed consent and surgical consent were obtained from all patients (WHO Trial Registration, ChiCTR-TNRC-08000193).

RPC isolation

The culture of RPCs was performed under Good Manufacturing Practice (GMP) conditions in the Cell Biology Therapy Center, Southwest Hospital, Third Military Medical University. This center has been awarded GMP certification and qualified for the production of RPCs.

Ocular tissues of 12- to 16-week-old aborted fetuses were collected from the embryonic tissue bank of the Department of Obstetrics, Southwest Hospital, according to the good tissue practice guidelines. Donors provided informed consent and were not compensated for the use of their terminated fetal tissue for research. Fetal neural retinas were cut into pieces, rinsed, digested at 37 °C for 20–30 min with Tryple (CTS, Gibco) and diluted by the addition of 3 ml medium (Ultraculture; Lonza)). The tissue was dissociated by gentle agitation for 10 s and the suspension was settled for 2 min. The supernatant containing the precursors was carefully decanted into a new tube with fresh medium, and the remaining pellet was discarded. The collected supernatant was centrifuged for 5 min at 2400 g and re-suspended in Ultraculture supplemented with 10 ng/ml human epithelial growth factor (EGF; Peprotech) and 20 ng/ml human basic fibroblast growth factor (bFGF; Peprotech) [14]. Cells were plated onto matrigel-coated tissue culture surfaces (Cellstart CTS, Invitrogen) and placed in an incubator for 120 min. The majority of non-neuronal cells adhered to the bottom of the plate, whereas neuronal progenitor cells remained in suspension such that the non-adherent suspension can be collected for primary culture and expansion [22]. Freshly purified RPCs were supplemented with Ultraculture, B27, N-2, 20 ng/ml human EGF, and 20 ng/ml human bFGF, placed on fibronectin-coated (10 μg/ml) plates and placed in an incubator (37 °C, 5% CO₂). RPCs at passage three were used for the following study. The viability of RPCs was determined using Trypan blue staining (0.4%; GIBCO).

Immunocytochemistry

The RPCs were plated onto glass cover slips. Primary antibodies were used to characterize the cells (PAX6, 1:100, Santa Cruz; CRX, 1:50, Santa Cruz; Nestin, 1:500, BD Bioscience; Sox2, 1:1000, Chemicon; GFAP, 1:1000, Chemicon). Cy3 (1:1000, Santa Cruz) was used as the secondary antibody and the slides subsequently imaged using a confocal microscope (Leica TCS NT, Leica Microsystems).

Flow cytometry

In brief, RPCs were prepared in a cell suspension, incubated with the primary antibodies (1:30 for Nestin, PAX6, SOX2, and GFAP) or isotype control (1:30; BioLegend), washed, and then incubated with fluorophore-conjugated secondary antibodies (1:30). Cells were analyzed on a fluorescence-activated cell sorting Calibur system (FACS, BD Biosciences, San Jose, CA, USA). The ratio of positive cells within the gated population was estimated based on comparison with species-specific isotype control. Ten thousand events/sample were collected, stored for analysis, and the experiments were repeated three times.

Real-time quantitative polymerase chain reaction (RT-qPCR) analysis

Total RNA was extracted from the RPCs by the RNeasy Mini Kit (Qiagen) and cDNA generated using the iScript cDNA Synthesis Kit (Bio-Rad) according to the manufacturer’s instructions. RT-qPCR) was carried out using a Power SYBR Green PCR Master Mix on the 7500 Real-time PCR System (Applied Biosystems). hESC line H1, which was a gift from Shanghai Institute of Biochemistry and Cell Biology, was used for comparison with RPCs. RT-qPCR assayed for the hESC markers Nanog and OCT4, and for the RPCs markers PAX-6, Six6, Crx, and recoverin. Relative gene expression was assayed in triplicate replicates normalized to the GAPDH signal present in each sample. The expression levels of cell markers detected in RPCs were normalized to that of an hESC sample which served as the zero set point.

Differentiation of RPCs into photoreceptors

Retinoic acid (10 μM; Sigma) was added into serum-free conditioned medium and the cells were cultured for 2 weeks in order to induce the RPCs to differentiate into mature photoreceptors [23]. Cells were then identified by the specific markers recoverin (1:1000, Chemicon) and rhodopsin (1:250, Chemicon). The cellular proliferating properties were examined by anti-Ki67and Ki67 (1:200, Abcam). Cy3-conjugated IgG was used as a secondary antibody.

Cell transplantation into RCS rats

RPCs were pre-labeled with the fluorescent marker CM-DiI (2 mg/ml; Invitrogen) prior to transplantation. For the efficacy study, RCS rats at 30 days old received an injection with RPCs (n = 12); 0.01 M phosphate-buffered saline (PBS) injections were used as a vehicle control (n = 6). The right eyes served as the treatment eyes, whereas the left eyes were untreated.

Rats were anesthetized by an intraperitoneal injection of a solution of ketamine (120 mg/kg) and xylazine (20 mg/kg). A scleral hole was created using a 30G needle allowing access to the space between the neural retinal layer and the retinal pigment epithelium layer. A glass micropipette carrying 5 μl of a RPC suspension (1 × 10⁵ cells) was inserted tangentially into the space beneath the degenerating photoreceptor layer at the superior retinal hemisphere. Fundus examination was performed immediately after surgery, and a successful injection was confirmed by a small subretinal fluid bleb. Cyclosporine A (200 mg/L) was given orally from day 1 until sacrifice.

Functional test after cell transplantation

Electroretinography (ERG) analysis was used to evaluate the improvement in retinal function after cell injections.

Three and six weeks following transplantation, animals were dark adapted for at least 12 h before the ERG test. Anesthesia was performed as above. Pupils were dilated using 1% tropicamide. The active gold lens electrode was placed on each cornea, and the reference and ground electrodes was respectively placed subcutaneously in the mid-frontal area of the head and the base of the tail. Light stimulation was delivered at –5 dB for the dark-adapted test, and all recordings were processed by software supplied by the manufacturer (Diagnosys LLC, MA). The amplitudes of a-waves were measured from the baseline to the cornea-negative peaks, and the amplitudes of b-waves were measured from the cornea-negative peak to the major cornea-positive peak.

Immunohistology

Rats were sacrificed at 6 weeks post-transplantation. The eyes were fixed in paraformaldehyde in PBS, infiltrated with sucrose, and then sectioned using a cryostat. The injected cells were preliminary identified by the fluorescent marker CM-DiI with fluorescence microscopy. Sections were washed in PBS three times to remove CM-DiI. Mouse anti-human mitochondria (1:200, Abcam) and rabbit anti-human recoverin (1:1000) or rhodopsin (1:250, Chemicon) were used as primary antibodies to detect the transplanted RPCs, and then sections were incubated in the secondary antibodies, Cy3-conjugated AffiniPure goat anti-mouse IgG (1:300) and FITC-conjugated AffiniPure goat anti-rabbit IgG (1:300).

We chose three rats to quantify the percentage recoverin/rhodopsin-positive cells among RPCs. From each rat, three random sections that containing the typical transplant areas were selected. The ratio of double-stained cells among the human mitochondrial-positive cells was considered as the photoreceptor cells differentiated from the grafted RPCs.

To compare the degree of outer nuclear layer (ONL) preservation between RPCs and vehicle groups, the thickness of the ONL was measured on the areas extending 100 μm either side of the injection site.

RPCs were also assessed for tumor formation in the retina. RPCs were injected (as above) into the space beneath the degenerating photoreceptor layer of P30 RCS rats (n = 36) and then examined 6 weeks post-transplantation. Hematoxylin-eosin staining was used to examine tumor formation in the injection area.

Patients

We enrolled eight patients diagnosed with rod-cone dystrophy on the basis of eye examinations, visual field testing, standard full-field fundus fluorescein angiography (FA), and flash (f)ERG according to the standards set by the International Society for Clinical Electrophysiology of Vision (ISCEV) [24]. Patients met the following inclusion criteria: (1) between 18 and 50 years of age; (2) best corrected visual acuity (BCVA) ≤ 20/400 in the operated eye, or a visual field of less than 20°, as assessed by Octopus 101 perimeter; and (3) the vision in the non-operated eye had to be better than the operated eye. Exclusion criteria included evidence of other eye disease such as a cataract that could compromise the interpretation of visual results; the inability to return for follow-up according to pre-planned schedule during the study; and history of intraocular surgery.

Surgical procedure for cell transplantation into retinitis pigmentosa patients

A standard three-port vitrectomy was performed and the vitreous body was removed from the inner limiting membrane of the retina. Using a 39G retinal hydrodissection cannula (Storz, USA), a minimally invasive retinotomy was performed temporally or superatemporally to the macula and near the arcade vessels. A RPC suspension (100 μl containing ~ 1 × 10⁶ cells) was slowly injected into the presumptive space thus creating a small retinal bleb (Additional file 1: Video S1). Cells were assessed prior to transplantation for microbial contaminants and endotoxin. Post-surgical treatment followed standard procedures for patients receiving three-port vitrectomy.

Clinical evaluation

Seven patients were followed for 24 months and one patient for 12 months. BCVA was measured three times at each visit using the Early Treatment Diabetic Retinopathy Study (ETDRS) chart. Data were then converted into logMAR (log of the minimum angle of resolution) scores according to the formula 1.1 + log₁₀ (designed distance/testing distance) – 0.02 × number of letters [2]. A high logMAR score indicates poor vision. Patients with only hand-motion vision were assigned a score that was one line lower than the largest printed line on the 4-m chart (<20/1600).

On each follow-up visit (1, 2, 3, 6, 9, 12, and 24 months post-transplantation), photographs and autofluorescence/fluorescein angiography of the fundus were performed using a Heidelberg HRA II system (Heidelberg Engineering GmbH, Germany). High-resolution optical coherence tomography (OCT; OCT-1000 System, Topcon) and Spectral Domain OCT (SD-OCT, Spectralis 3 Mode OCT, Heidelberg Engineering) were used to evaluate retinal structure. Bilateral full-field ERGs were recorded using a Roland electrophysiology system (RETIscan, Roland Consult, Germany) with ERG-jet contact lens electrodes. For the ERG analyses, the pupils were dilated with 1% tropicamide and the patient dark-adapted for 30 min. ISCEV standard dark-adapted and light-adapted ERGs were recorded.

Pupillary light reflexes in the dark-adapted (40 min) state were evaluated using a custom-built computerized pupillometer and a modified commercial spherical Ganzfeld according to the method of Aleman et al. [25]. Five continuous 200-ms blue stimuli with intensities ranging from –2.8 to 0.85 log scot-cd/m² and six white stimuli ranging from –1.5 to 2 log scot-cd/m² were applied to elicit a transient light reflex; each stimulus (from low to high) was followed by a 15-s dark recovery period [26]. It was difficult to analyze the amplitude changes for individual pupillary reflexes due to large amplitude variations elicited by the same stimulus. A response criterion of 0.3 mm was used to define a response threshold. The threshold values were converted to ranked data. A one-level improvement corresponded to a one-level decrease in the threshold.

Statistical analysis

Data are given as the mean ± SD. Comparisons were made using a two-tailed paired t test for visual acuity. The treatment effect was compared to the baseline condition and that in the follow-up time points. Differences in BCVA (logMAR) were obtained for each patient at 0, 1, 2, 3, 6, 9, 12, and 24 months post-transplantation. The differences following treatment at each time point were normalized to the baseline measurement obtained at month 0 in order to perform comparisons across patients. A chi-square test was used to assess differences in the pupillary light reflexes between the operated and non-operated eye. All statistical tests were considered significant if P ≤ 0.05.

Human fetal-derived RPC transplants in RCS rats

RPCs after three passages were characterized by checking the expression of Nestin (98%), Pax6 (96.6%), and Sox2 (78%) using immunocytochemistry or flow cytometry (Fig. 1a and b). There were limited glial cells in the population of PRCs (0.1% GFAP⁺ cells). Gene expression analysis confirmed that the typical early eyecup transcription factor genes Pax6 and Six6 (Fig. 1c) increased 5–10 fold compared to that seen in hESC cultures; the expression of photoreceptor precursor markers, such as recoverin and CRX, was much higher in the RPCs. In contrast, the expression of the pluripotency markers NANOG and OCT4 clearly decreased by 5–10 times compared to hESCs. RPCs were also able to differentiate and express photoreceptor phenotypes (recoverin and rhodopsin) in vitro following treatment with retinoic acid, and lost proliferative properties without Ki67 staining (Fig. 1e). Since RPCs would be transplanted into degenerated retina in retinitis pigmentosa patients, these cells were extensively tested for animal and human pathogens. The final RPC product had normal female (46 XX) karyotype, confirming that the cells were free of microbial contaminants (Fig. 1d).

Six week following human RPCs transplantation, DiI-labeled cells could be readily identified within RCS retinal sections and, in Fig. 2a, it is clear that transplanted cells had spread across a broad area of the region previously occupied by the photoreceptors. Moreover, immunofluorescence staining confirmed that these cells exclusively expressed human mitochondria, a specific marker of the human species. Most RPCs integrated in the host ONL. Some cells co-expressed recoverin or rhodopsin (Fig. 2b and c). We observed approximately 20.7 ± 3.1% recoverin-labeled and 9.7 ± 1.5% rhodopsin-labeled RPCs in the recipient ONL in each randomized section, indicating the expression of a photoreceptor phenotype and possible photoreceptor differentiation.

Statistical analysis of ONL thickness indicated that transplants were associated with a significantly thicker ONL compared with that in the PBS group (37.2 ± 2.8 μm vs 18.4 ± 2.0 μm, P < 0.05; Fig. 2d). Analysis of the recorded ERGs at 3 and 6 weeks post-transplant indicated that the b-wave amplitudes were higher compared to those of the PBS group (P < 0.05; Fig. 2e and f). These findings showed that retinal function was profoundly improved after RPC transplantation, which corresponds to the morphological results by ONL thickness measurements.

We then tested the risk of tumor formation by injecting RPCs into the degenerating retinas. Cellular survival was observed in 32 eyes from 36 RCS rats; tumors were not observed in any retinal sections, suggesting the safety of human RPC transplantation in retinitis pigmentosa patients.

Clinical study of fetal-derived RPC transplantation in retinitis pigmentosa patients

Preoperatively, the visual acuity of the prospective transplanted eye was 1.37 ± 0.34 logMAR in the eight patients and is equivalent to ~ 20/500 on a Snellen’s vision chart. Vitrectomy and subretinal transplantation proceeded without complications, such as iatrogenic retinal tears, cataracts, or endophthalmitis (Table 1 and Additional file 9: Video S1). All postoperative retinal examinations on day 1 were unremarkable, apart from the formation of the retinal bleb containing the transplanted cells. Clinical examination on day 7 revealed that retinal detachments had not occurred and the OCT analysis clearly showed a thick region of transplanted cells beneath the neural retinal layer (a typical example is shown in Fig. 3b). Although reduced in thickness, the layer of RPCs was still clearly present in OCT scanning 1 month after surgery (Fig. 3c). The reduced thickness at the transplant site suggested that the grafted cells may have migrated within the subretinal space or integrated into other layers of the host retina, in a similar fashion to that seen within the RCS rat retinas. One and two years post-transplantation, it was no longer possible to unequivocally define the region of injected cells using the OCT methodology (Fig. 4 and Additional files 2–7: Figures S2–S7). In some cases the region of the injection site could be defined by the presence of small retinal scarring, characterized by locally thickened retina with OCT scanning. Autofluorescence provided useful information on conditions where the health of the RPE played a key role; areas of hypo-autofluorescence indicated missing or dead RPE cells [27]. Autofluorescence imaging showed a limited hypo-autofluorescence area beneath the corresponding retinal injection site, a sign of local RPE disruption (Fig. 4c’). The restricted area of these regions indicated that our surgery was safe and minimally invasive.

			BCVA (logMAR)		Donor cells
Patients	Age (years)	Sex	OS	OD	Gestational age (weeks)	Viability
1	42	M	1.56	1.6	14	>98%
2	33	F	1.30	1.3	15	>98%
3	34	M	1.00	0.8	12	90–93%
4	19	M	0.35	0.3	12	95%
5	46	F	1.52	0.9	15	96%
6	19	F	1.30	0.4	12	94%
7	53	F	2.00	1.8	12	95%
8	38	F	0.92	1.0	12	93%

BCVA best corrected visual acuity, F female, M, male, MAR minimum angle of resolution, OD right eye without treatment, OS left eye with transplant

Fluorescein angiography indicated that our procedure did not lead to changes in vascular leakage in the region of the macular in eight patients. Distribution of autofluorescence also showed no absent autofluorescence in the macular area compared to the baseline (Fig. 4c’ and c”), indicating that our injected cells did not lead to further retinal degeneration or oxidative injury to the remaining functional RPE cells. OCT scanning throughout the 24-month follow-up period did not detect the presence of inflammation or cystoid macular edema (Fig. 4 and Additional files 2–7: Figures S2–S7), and none of the patients developed any sign of immunological rejection in the fundus. However, one patient formed a macular membrane (Additional file 1: Figure S1). Given these results, medication for systemic or intraocular immunosuppression was not administered.

Visual acuity in the RPC-treated eyes showed a significant improvement for grouped data compared to the baseline (P < 0.05) between 2 and 6 months after surgery (Fig. 5, Table 2), but acuity then declined so that overall no differences were seen by 24 months. When individual patients were examined, BCVA improved in five eyes, but remained stable in three eyes during the 12 month follow-up; at 24 months, improved vision was only seen in one eye.

			Follow-up month
Patient	Eyes	Baseline	1	2	3	6	9	12	24
1	OS	1.56	1.52	1.30	1.30	1.30	1.52	1.80	1.90
	OD	1.58	1.52	1.34	1.52	1.52	1.60	1.60	1.30
2	OS	1.30	1.30	1.20	1.30	1.20	1.30	1.30	1.32
	OD	1.30	1.30	1.30	1.20	1.30	1.30	1.30	1.16
3	OS	1.00	1.00	1.00	0.80	0.80	0.80	0.80	1.00
	OD	0.80	0.80	0.80	0.90	0.90	0.90	0.90	0.90
4	OS	0.60	0.60	0.56	0.54	0.54	0.54	0.56	1.34
	OD	0.34	0.34	0.40	0.42	0.50	0.50	0.54	1.20
5	OS	1.52	1.00	1.00	1.30	1.52	1.60	1.60	1.90
	OD	0.90	0.90	0.90	0.90	0.90	0.90	0.90	1.18
6	OS	1.30	1.00	1.00	1.13	1.26	1.30	1.30	1.15
	OD	0.40	0.30	0.30	0.50	0.40	0.50	0.50	0.70
7	OS	2.00	2.00	2.00	2.00	1.90	2.00	2.00	2.00
	OD	1.90	1.90	1.90	1.90	1.70	2.00	1.90	1.90
8	OS	0.92	0.90	0.90	0.92	0.92	0.92	0.90
	OD	0.80	0.80	0.80	0.80	0.80	0.80	0.80

MAR minimum angle of resolution, OD right eye without treatment, OS left eye with transplant

The recorded signal during global ERG measurements was too small and not clearly distinguishable either before or after RPC transplantation, indicating the poor visual function of the patients prior to the study. Similarly, visual field tests were also unreliable for this reason. Therefore, we developed a computerized pupillometry to assess photoreceptor dysfunction. The thresholds for blue and white light-induced pupillary light reflexes were recorded. The threshold to blue light improved in four subjects (patients 1, 3, 4, and 7) between 3 and 6 months post-transplantation (Additional file 8: Figure S8A, C, D, and G), remained stable in three others (patients 2, 6, and 8; Additional file 8: Figure S8B, F, and H), but declined in one patient (patient 5; Additional file 8: Figure S8E). The threshold to white light improved in patients 1, 3, and 4 (Additional file 8: Figure S8A’, C’, and D’) and remained stable in the other five patients. These findings indicated that, compared to the preoperative baseline, the retina became more photosensitive during the 6-month post-transplantation period in patients 1, 3, and 4, but this was not sustained and threshold levels declined to the baseline at 12 months. In the other five patients, three maintained preoperative baseline levels, while two patients showed a decline in the light reflex.